Tips for cold-starting short video community content!

The short video community is dominated by content. The differences between the tool modules of various products are not significant. Differentiating the content is the fundamental to forming the tone of the community. The so-called tone can be understood as "the collective preference of the community for similar content in a specific scenario."

Back to the spring two years ago, TikTok topped the App Store download list and remained there for many times. Kuaishou, the industry's big brother with hundreds of millions of users, has begun to face external doubts about its lack of growth.

It seems that overnight, the streets and alleys were filled with the brainwashing rhythm of Haicao Haicao and C哩C哩; the industry was endlessly brainstorming about "centralization" or "decentralization"; and the official accounts were filled with heated discussions on "who will be the end of the short video community". The entire industry is shrouded in anxiety and excitement - everyone wants to get a piece of the pie, as if it would be too late if they don’t get in.

It was in this environment that a short video community product was quietly launched. As a community operator, facing the incubation of a new short video community from 0 to 1, the cold start pressure is huge and we don’t have much time left.

Operational goals of cold start communities

After the product was successfully launched, the first question facing us was what kind of operational goals to set during the cold start phase.

At this time, many operations colleagues are eager to increase several so-called "core indicators" such as the amount of community videos, the number of community creators, new downloads, and the next-day retention of new users. With very limited operational resources, we started to launch production plans and recruit content creators. We frequently organized brainstorming sessions to explore operational topics and activities that could be quickly launched online. We also participated in UI and product discussions, striving to help the product quickly and at low cost in the tool module from an operational perspective.

However, these efforts often have little effect and a low input-output ratio. The high retention rate at the beginning of the product launch due to the freshness and ease of use of the tool module will gradually lose its advantage after the community is launched.

Reflection: As front-line operators, it is easy for us to fall into a short-sighted vicious circle. Seeing the mountains of operational tasks at hand, I feel like I am picking up sesame seeds and losing watermelons. No matter what position you hold within a product, this is a manifestation of a lack of long-term thinking about the product's operating rhythm and the overall project goals.

As mentioned earlier, content differentiation is important to a nascent content community. Therefore, in the cold start phase, if you want to establish scientific and effective operational indicators, you should first clarify a clear content operation goal. Simply put, it is to quickly find the content that users prefer and start building an effective mechanism to provide this content from that moment on.

Therefore, the core indicators of community operations at this stage should focus on indicators such as the average number of video plays per person, the average playback time per person, and the sharing rate of a single video, which can intuitively and quickly reflect user content preferences. Only by taking this first step well can a truly vibrant content community be established. Only then will macro data such as retention and download volume have the vitality to develop healthily. On the contrary, if this cannot be achieved, there will be no such thing as differentiation of community content.

Find out what the community residents prefer

After clarifying the goals of content operations, we need to start paying attention to the impact of community content on user behavior. The goal is to understand the content preferences of a large number of users in the community. The logic is very simple: in the initial stage of a community, if there are few resources and the direction is not clear enough, the word-of-mouth effect brought about by the content on the site is the best way to trigger user self-propagation.

Many products that focus on video editing functions face the problem of scarcity of high-quality content in the early stages of their launch. Solving the content source problem has become a top priority for operations colleagues. At this stage, in order to streamline the main steps of content screening, content supply, content recommendation and review and summary, almost every community cold start will adopt the same method: official involvement.

By filtering the original content of users on the site and the popular content on the entire network, the target content is first roughly divided into several common dimensions such as channels and scenarios. Next, we will release various types of content in a certain proportion on a daily or weekly basis, and by observing and comparing the content consumption data, we can quickly understand the community residents' preference for different content.

It must be said that this process is cumbersome and arduous. When the community user base is small, it requires a high time cost for content operators. But despite this, this link is essential for the cold start of the content community.

If a novice operator wants to understand the rules of the game in a community, the best option is to join the community and become a member. Get first-hand and intuitive experience by operating your own account. Similarly, the various contents that operations students put into the community are just a repeated test by a novice up-host to find potential fans of his account. If the essence of operations is to find a better way to serve users for the product, then only by becoming a user can you better serve users.

With only 100,000 existing users, the content testing process lasted approximately 2-3 weeks.

This method can help operations staff quickly grasp a series of content consumption data, so that the composition of future community content gradually becomes clear. The focus of operations personnel is on basic indicators such as the number of single video plays, effective play rate, click-through rate, and interaction rate.

Operations colleagues will find that some content that is widely praised on the top platforms may have a mediocre response within a brand new community; while other content that was not favored in the early stage of release may have outstanding consumption data. This reflects that the leading platforms cannot satisfy the content preferences of all users, and new content communities will always have opportunities. Despite this, anchoring cost remains an important factor affecting user migration.

Reflection: The target videos come from different sources. It is unfair to put user-generated content, popular content on the entire network, and content produced by experts in one pool for direct competition. The data obtained may also blind the observation of operators and affect their judgment of the content.

In fact, no matter how tedious or detailed the operational work during the cold start test phase is, it is not excessive. A moment of inconsideration may lead to a butterfly effect. Therefore, a more scientific, detailed and data-indicator-focused content delivery plan should be developed at the beginning of the experiment.

When observing video consumption data, there is one thing that is easily overlooked by operators: splitting up these individual data and comparing them horizontally with the overall consumption data and the consumption data of similar products.

A better way to experiment would be to compare data such as video play times, click-through rates, and interaction rates with the overall average, paying more attention to the differences between each individual data and the overall data. This is like after the monthly exam for senior high school students, instead of focusing on the improvement of their grades, it is more important to focus on their ranking in the school, city, and province.

If we find the preference, does that mean we have found the tone of the community?

As mentioned before, the so-called tone can be understood as "the collective preference of the community for similar content in a specific scenario", which includes the 50,000 questions when making any product.

Back to the perspective of the cold start of the content community, content advantage is the foundation of all operational methods. In order to create a community atmosphere and tone, the first thing to do is to improve the quality of videos on the site and form the so-called content advantage. Only the content itself can help the community find its own tone and atmosphere.

For operations colleagues, "quality" includes both content quality and content inventory.

At this stage, new communities begin to access recommendation algorithms. Too small a content stock cannot provide the basic conditions for the deployment of algorithm strategies. Therefore, at this time, operators must start looking for ways to quickly attract new users and fundamentally expand the community's user base.

At this stage, operators can start launching lightweight activities that meet user expectations and raise interactive topics; and based on the basic portrait of the user population obtained in the previous stage, make small and detailed precision deliveries in some channels.

With this wave of operations, the overall data will generally have a phased improvement. When the promotion effect is good, the product is likely to usher in its first small peak of rapid growth. The challenge facing operators at this time becomes how to satisfy and retain these users.

The huge influx of traffic in the short term is a double-edged sword. On the one hand, it brings greater potential creative capabilities and a more diverse consumer group; but on the other hand, high-quality content will quickly become more scarce.

Before this group of new users leaves, operations staff must quickly capture some of them with high-quality content and keep the customer acquisition cost within an acceptable range. Only in this way will there be more opportunities for subsequent operations and promotions.

During the user content preference testing phase mentioned above, it is mainly the operations staff who lead the content supply. At this stage, faced with a user base that is several times larger than before, continuing to rely on operators to produce content themselves is not only a drop in the bucket, but the community will have no future at all. Finding producers and finding a stable source of content production became the focus of operational work for a long time from this moment on.

In the process of exploring content production mechanisms and stimulating user production, the most critical link in the establishment of a content community has surfaced. As an operator, the first thing you need to do is to tell users which content is encouraged to be produced; secondly, you need to incentivize content production in a targeted manner; finally, you need to introduce user (producer) operations at the appropriate time and incorporate them throughout the entire community development.

By reviewing and recommending content, the official is sending a clear message to users: what kind of content is more likely to gain exposure, and what kind of content does not conform to the basic atmosphere of this community. The work of transmitting this information is initially done manually by the operations staff. At this stage, it will gradually transform into [manual recommendation] + [algorithm recommendation].

Reflection: In this process, the official should let users know clearly what kind of content the community needs. This is not only the achievement of psychological expectations of both parties, but also the first step in forming a community atmosphere.

If in this process, the authorities put some kind of mental cleanliness (such as what they want to see and what they don't want to see) above objective user preferences; or are wavering and ambiguous, or blindly increase the weight when the recommendation algorithm is not yet perfect, disrupting the operational plan, these will have long-term consequences.

Author: The core of the food is finished

Source: The food is finished